Sarah Friar (세라 프라이어), OpenAI's chief financial officer, said the company could slow development of cutting-edge models if necessary for AI safety. She publicly signalled that safety and alignment issues could take priority over the race for performance.

On Sept. 15 (local time), CNBC reported that Friar said in an interview with Jim Cramer (짐 크레이머) that "we need to take safety seriously." She said if frontier development needs to be regulated or slowed, she would follow the researchers' judgement. She added that as CFO she would adjust investment and business plans accordingly.

The comments come as debate over AI safety spreads again. Former Anthropic researcher Jacob Coxon (제이컵 콕슨) recently left the company and criticised major AI firms for advancing technology too quickly, warning that advanced AI could pose a fatal risk to humanity. This is Coxon's personal risk outlook, and the likelihood of it actually occurring has not been proven.

Dario Amodei (다리오 아모데이), Anthropic's chief executive, also made a proposal on Sept. 12 aimed at regulating the pace at the frontier. Amodei said independent evaluation bodies should be given access sufficient to continuously examine internal lab systems and that major AI companies should establish common safety standards. OpenAI CEO Sam Altman and Elon Musk also expressed support for that direction.

The safety debate is also affecting OpenAI's initial public offering schedule. Altman recently told Fortune that a 2026 listing would not be appropriate given safety issues, ruling out the possibility of an IPO this year. Friar previously told employees she expected a listing in 2027.

Friar's remarks show that AI safety issues are expanding from technical discussions in research departments into management variables that could influence investment, product launches and listing schedules.